To me that just seems like another reason, if you're someone backing multiple startups exposed to the overall health of the ecosystem, to point people to different options so you don't have a single point of failure.

Surely there are enough banks and bankers around that you could have half a dozen or more specialists? It's knowledge work, not something needing billions in physical infrastructure investment.

The NY Fed looked at lending specialization actually improved Bank lending performance https://www.newyorkfed.org/medialibrary/media/research/staff...

This has to do with deposit specialization which I don’t know if it’s been studied, but you could see the same arguments applying to.

If you're an individual startup, you don't care about the overall system, just maximizing your own. So that efficiency seems great.

But if you're a VC with a broad portfolio, you're much more exposed to the single-point-of-failure risks that come with increased efficiency.

"Give up some efficiency for system robustness and long-term-health" is a no-brainer for mature engineering orgs, but in this financialized world nobody can see the longer term big picture